The goal of this project is to develop new prodrugs that become activated to cytotoxic anticancer drugs exclusively in solid tumors. These nontoxic prodrugs will be enzymatically activated to cytotoxic agents by the nonpathogenic obligate anaerobe C. sporogenes genetically engineered to express the prodrug-activating enzymes. When spores of these genetically engineered C. sporogenes are injected intravenously the spores germinate exclusively in the hypoxic necrotic areas of tumors and produce high levels of the enzyme exclusively in these regions. Such hypoxic/necrotic regions are extremely common in human solid tumors and provide the basis for the selective expression of the prodrug-activating enzymes in the tumors. When a nontoxic prodrug is administered systemically following the intravenous injection of the genetically engineered bacterial spores, conversion to the cytotoxic metabolite occurs exclusively in the tumor. We have demonstrated proof-of-principle of this CDEPT (clostridia-directed enzyme prodrug therapy) approach with the E. coli derived enzyme cytosine deaminase which converts the nontoxic drug 5-fluorocytosine to the anticancer drug 5-fluorouracil, and with E. coli derived nitroreductase which converts the dinitrobenzamide prodrug CB 1954 to a toxic bifunctional alkylating agent. We propose to evaluate newly synthesized prodrugs with three bacterial enzymes: nitroreductase, beta-glucuronidase, and beta-galactosidase in this system. For each of these enzymes we will test prodrugs designed and synthesized for their conversion to the toxic drug, for their ability to diffuse through a multicellular layer and for their antitumor efficacy, both alone and in combination with single doses and fractionated irradiation. The combination with radiation will test the hypothesis that there will be greater cell kill by the activated prodrugs of the hypoxic compared to the aerobic tumor cells. We will also combine this approach with two vascular-targeting drugs, the TNF-inducing drug DMXAA and the tubulin-binding agent ZD6126, both of which decrease blood flow and increase necrosis in experimental and human tumors. We hypothesize that these vascular targeting agents will sensitize very small tumors (lung metastases) to CDEPT because of their ability to induce necrosis in such tumors.